A comparative study of four input devices for desktop virtual walkthroughs

(1)

Publisher’s version / Version de l'éditeur:

Computers in Human Behavior, 27, 6, pp. 2186-2191, 2011-07-19

READ THESE TERMS AND CONDITIONS CAREFULLY BEFORE USING THIS WEBSITE.

https://nrc-publications.canada.ca/eng/copyright

Vous avez des questions? Nous pouvons vous aider. Pour communiquer directement avec un auteur, consultez la première page de la revue dans laquelle son article a été publié afin de trouver ses coordonnées. Si vous n’arrivez pas à les repérer, communiquez avec nous à PublicationsArchive-ArchivesPublications@nrc-cnrc.gc.ca.

Questions? Contact the NRC Publications Archive team at

PublicationsArchive-ArchivesPublications@nrc-cnrc.gc.ca. If you wish to email the authors directly, please see the first page of the publication for their contact information.

NRC Publications Archive

Archives des publications du CNRC

This publication could be one of several versions: author’s original, accepted manuscript or the publisher’s version. / La version de cette publication peut être l’une des suivantes : la version prépublication de l’auteur, la version acceptée du manuscrit ou la version de l’éditeur.

For the publisher’s version, please access the DOI link below./ Pour consulter la version de l’éditeur, utilisez le lien DOI ci-dessous.

https://doi.org/10.1016/j.chb.2011.06.014

Access and use of this website and the material on it are subject to the Terms and Conditions set forth at

A comparative study of four input devices for desktop virtual

walkthroughs

Lapointe, J.-F.; Savard, P.; Vinson, N. G.

https://publications-cnrc.canada.ca/fra/droits

L’accès à ce site Web et l’utilisation de son contenu sont assujettis aux conditions présentées dans le site LISEZ CES CONDITIONS ATTENTIVEMENT AVANT D’UTILISER CE SITE WEB.

NRC Publications Record / Notice d'Archives des publications de CNRC:

https://nrc-publications.canada.ca/eng/view/object/?id=ca17fe75-d39e-4ed9-a7be-7a1c6c1b8cb3

https://publications-cnrc.canada.ca/fra/voir/objet/?id=ca17fe75-d39e-4ed9-a7be-7a1c6c1b8cb3

(2)

Other uses, including reproduction and distribution, or selling or

licensing copies, or posting to personal, institutional or third party

websites are prohibited.

In most cases authors are permitted to post their version of the

article (e.g. in Word or Tex form) to their personal website or

institutional repository. Authors requiring further information

regarding Elsevier’s archiving and manuscript policies are

encouraged to visit:

(3)

Author's personal copy

A comparative study of four input devices for desktop virtual walkthroughs

J.-F. Lapointe

⇑

_{, P. Savard, N.G. Vinson}

Institute for Information Technology, National Research Council of Canada, 1200 Montreal Road, Building, M-50, Ottawa, Ontario, Canada K1A 0R6

a r t i c l e

i n f o

Article history:

Available online 19 July 2011

Keywords:

Virtual walkthroughs Input devices

Desktop virtual environments Control devices

Navigation Travel

a b s t r a c t

This paper presents the results of an experiment measuring the effect of four different input devices on overall task performance for desktop virtual walkthroughs. The input devices tested are: a keyboard, a mouse, a joystick and a gamepad. The results indicate that the participants completed the tasks in signif-icantly less time and distance travelled with the mouse than with the three other input devices. The use of the mouse also significantly reduced the number of collisions, while the use of the gamepad resulted in significantly more collisions.

1. Introduction

Virtual walkthrough is a common travel metaphor used for viewpoint control in virtual environments (VEs) (Brooks, 1986; Usoh et al., 1999). Today, a large proportion of users perform walk-throughs on desktop VEs (i.e. displayed using a monitor) as op-posed to immersive displays (e.g. HMDs, CAVEs) (Ruddle, Payne, & Jones, 1997; Santos et al., 2009). Virtual walkthroughs can be di-vided into a set of cognitive tasks and a set of physical tasks that a person performs to get from one location to another (Bowman, Kruijff, LaViola, & Poupyrev, 2005; Sebok, Nystad, & Helgar, 2004). While many studies have investigated the cognitive aspects (orientation and wayfinding) of walkthroughs (e.g.Ruddle, 1999; Ruddle, Payne, & Jones, 1998; Ruddle et al., 1997) we are focusing on travelling itself, which is the physical tasks that allow a naviga-tor to move from one location to another by using the input de-vices (a.k.a. control dede-vices) of the user interface (Bowman et al., 2005; Sebok et al., 2004). The input device design space for virtual walkthroughs is quite large given the possible combinations of commands that can be issued from a number of input devices. However, very little of this design space has been investigated empirically. As with other user interfaces, empirical evaluation of possible interaction techniques is important to improve our under-standing and to develop theoretical models and to assist designers in building more usable systems (Fröhlich, Hochstrate, Kulik, & Huckauf, 2006; Lampton et al., 1994; Lindeman, 2006; Newell & Card, 1985).

In virtual walkthroughs, the number of Degrees Of Freedom (hereafter ‘‘DOF’’) of movement varies between 2 and 4, out of a

potential 6. Three of the six DOFs involve translation: forward/ back, left/right, and up/down (which is typically not needed in vir-tual walkthroughs). The other three DOFs involve rotation: rotation around the axis perpendicular to the travel plane (yaw axis), and rotation around each of the axes forming the travel plane (pitch and roll axes), though the last one is not typically needed in virtual walkthroughs. Each DOF can potentially be controlled by a differ-ent input device and a differdiffer-ent input command from that device. Moreover, several travel techniques are possible by using one or two-hands.

To date, the research shows that, in the case of 3D interface, there is still not an input device that demonstrates its superiority for accomplishing basic 3D tasks such as navigation, manipulation and selection. This contrasts with the case of the 2D graphical interface where the computer mouse established itself as the de facto standard input device (Zhai, 1998).

In the case of maze travelling tasks, research indicates that bimanual travel control is quite feasible and even outperforms the status quo mouse-mapping interface (Zhai, Kandogan, Smith, & Selker, 1999). Also, two studies did not find significant differ-ences in performance when travelling with a user interface using two or three DOFs, when the third DOF is a strafe movement (i.e. a side translation of the viewpoint) (Lapointe & Savard, 2009; Lapointe & Vinson, 2002). Finally, another study suggests that the use of either velocity or position control techniques for viewpoint orientation does not have a large effect on travel performance (Lapointe & Savard, 2007).

For desktop walkthroughs, commonly found control interfaces in video games use a keyboard and mouse combination for PC games and two joysticks (on a gamepad) for console games. The use of these specific combinations appears to be driven by avail-ability of the input devices and is not backed by published scien-tific analysis of the performance or usability of these interfaces.

⇑ Corresponding author.

E-mail addresses:jean-francois.lapointe@nrc-cnrc.gc.ca(J.-F. Lapointe), pascal.-savard@nrc-cnrc.gc.ca(P. Savard),norm.vinson@nrc-cnrc.gc.ca(N.G. Vinson).

Computers in Human Behavior 27 (2011) 2186–2191

Contents lists available atScienceDirect

Computers in Human Behavior

(4)

To place the design and use of walkthrough interfaces on a more empirical footing, and to contribute to the exploration of the de-sign space, we conducted an experiment in which we compared the user’s performance and preference when using four different input devices, using either two or three DOF to travel in desktop virtual walkthroughs. We employed four input devices that are commonly used for desktop virtual walkthroughs: a keyboard, a mouse, a joystick and a gamepad.

Hence, this paper describes the results of an experiment that aims to characterize the usability benefits of those control devices for doing desktop virtual walkthroughs.

2. Method

To evaluate the different travel techniques, we used a maze-like virtual world made of a complex trail offering an open view, so that users could always look around and/or keep an eye on the end point of the maze while they travelled (Fig. 1). This open-view maze offered more incentives to use all the available DOF than a traditional maze environment. A traditional maze has high walls in which only views along the center line of a corridor offer inter-esting visual cues to spatially orient the users.

2.1. Participants

Of the 34 participants who began the experiment, eight were unable to complete it due to cybersickness, and two were outliers, exceeding all participants’ average trial completion time by more than two standard deviations. This left 24 participants that

corre-sponded to the 24 interface presentation order permutations, pro-ducing a properly counterbalanced experiment.

Fifteen of the 24 unpaid participants were male, and nine were female. Twenty of them were right handed, three were left handed and one was ambidextrous. They were all computer literate, had at least a high-school-level education, and an age distribution of 18/ 58/36 (min/max/average). They all had normal or corrected-to-normal vision.

2.2. Task

Participants had to complete a primed search task, where they knew in advance where the target (end point) was positioned (Bowman et al., 2005). They were instructed to travel from the start point to the end point in the shortest time possible (Fig. 2).

2.3. System

The system included a color desktop monitor with a diagonal size of 54 cm and a resolution of 1600 1200 pixels. The frame rate was 60 Hz, with system latency less than 120 ms. The partic-ipant’s viewing distance was 70 cm. The avatar had a radius of 0.25 m, a viewing height of 1.8 m, and a Field Of View (FOV) of 75° 60° (H V). The virtual trail was 2 m wide with walls that were 3 m high and obstacles that were 1 m high, thus allowing the participants to see over them. The corridor was 164.5 m long (measured along the center line), with 15 turns to the right and 15 to the left (seeFig. 2).

2.4. Travel techniques

Participants used four different travel techniques for their vir-tual walkthroughs. By technique, we mean a particular combina-tion of an input device, a number of hands used to control the device, a number of DOFs, and a device-to-avatar movement map-ping. Each technique moved the avatar and viewpoint together, such that the avatar’s viewpoint always faced in the same direction as the avatar’s forward motion. For all travel techniques, collisions between the avatar and objects of the scene (obstacles or walls) were slippery, so that the avatar did not get stuck in a corner. In all cases, a rate control algorithm controlled the translations of the viewpoint in the VE. A pilot study helped us select speed con-trol mechanisms for each travel technique to optimize user performance.

The first travel technique involved using a standard 104 key keyboard from Dell (Dell SK-8135) to control 3 DOFs (Fig. 3). The four directional arrow keys had two translation DOFs: fore/aft and left/right without any rotation (a.k.a. strafe). The 4 and 6 keys of the numerical keypad produced a rotation along the horizontal plane, around the vertical axis (a.k.a. yaw). Speed was constant at 5 m/s for translations and 180°/s for rotations.

Fig. 1. The walkthrough virtual maze environment.

Ending point Starting point

(5)

Author's personal copy

The second travel technique is based on a Microsoft Laser Mouse 6000 mouse (Fig. 4). Here, the lateral, side-to-side move-ment of the mouse rotated the viewpoint along the horizontal plane, around the vertical axis, (a.k.a. yaw). Rotation speed was determined by an algorithm with a linear function gain of 25°/ cm (with mouse acceleration disabled). The left and right buttons controlled the fore/aft movement at a constant speed of 5 m/s. Thus, the mouse travel technique had only two DOFs: one transla-tion and one rotatransla-tion.

The third travel technique involved a 3-DOF joystick as the in-put device (Fig. 5). A lateral tilt of the joystick resulted in a lateral displacement of the viewpoint (strafe motion) while forward/back-ward joystick tilt produce a fore/aft movement of the viewpoint. Finally, a twist of the joystick’s handle rotated the viewpoint on the horizontal plane, around the vertical axis (yaw). Translation speed was controlled through a linear function gain up to a maxi-mum speed of 5 m/s. Rotational speed was a constant 180°/s.

The fourth and final travel technique employed a Logitech Dual Action gamepad with two mini joysticks to control 2 translation

DOFs and one rotation DOF. As with the 3D joystick, a lateral tilt of the left joystick produced a lateral movement (strafe) of the viewpoint. A foreword/backward tilt of the left joystick produced the fore/aft movement of the viewpoint. Unlike the 3D joystick, the horizontal plane rotation, around the vertical axis, (yaw) was controlled by the second joystick. Translation speed had propor-tional velocity control with a linear function gain, to a maximum of 5 m/s. The rotation speed was a constant 180°/s (seeFig. 6).

2.5. Design

The independent variables were the four travel techniques described earlier and the dependent variables were the task completion time, the total traveled distance and the number of col-lisions. A collision is detected when a movement of the avatar to-wards an obstacle causes a contact. Collisions are detected at each frame, i.e. 60 times/s, thus implying a maximum of 60 collisions/s. We used a within-subject design and trials were blocked by travel technique. We counterbalanced the order in which participants used each travel technique to minimize skill transfer effects. For each travel technique (each block), the participants received brief explanations by the experimenter, followed by two practices and five trials.

2.6. Procedure

Participants read the instructions and completed a consent form along with a background questionnaire. They were then seated in front of the system and told to begin the experiment.

The instructions were displayed on-screen before each trial. A 3 s audio countdown preceded each trial. The trials ended auto-matically when the participants reached the end point. Once the trials completed for all travel techniques, each participant was in-vited to complete a post-test survey to rate each travel technique on the ease-of-use, fatigue, accuracy, speed and preference.

That survey builds on previous work in the field to measure ease-of-use and fatigue (Zhai, 1993) by using a standard five-point Likert scale to measure each item. The motivation for conducting the questionnaire survey is to acquire valuable feedback from the participants.

3. Results

3.1. Quantitative results

We first determined the effect of practice on performance. This allowed us to exclude practice effects from our analyses of the travel technique performance effects.Fig. 7shows that

practice-Fig. 4. The mouse travel technique.

yaw

stra

_fe

fo

re

/a

ft

Fig. 5. The joystick travel technique.

Fig. 6. The gamepad travel technique. Fig. 7. Effect of practice on mean completion time. 2188 J.-F. Lapointe et al. / Computers in Human Behavior 27 (2011) 2186–2191

(6)

related performance improvements on completion time disap-peared after a few trials. Consequently, we restricted our travel technique performance analyses to the last three trials in each block.

Fig. 8illustrates the mean task completion times for the 4 travel techniques (average of the last three trials). An analysis of variance (ANOVA) with the pseudo-F test was significant, F(3, 69) = 10.30,

p < 0.001.

A Duncan’s multiple range test (Duncan, 1955) reveals that the only significant difference is between the mouse travel technique and each of the three other ones.

The mean completion time was between 20% and 26% higher when using the keyboard, joystick or gamepad than when using the mouse. The effect size (using Cohen’s method (Cohen, 1988) is illustrated inTable 1.

We can see that there is a large effect size (i.e. >0.8) between the mouse and each of the three other interfaces.

Fig. 9illustrates the results for the three interfaces for the trav-elled distance (average of the last three trials). An analysis of var-iance (ANOVA) with the pseudo-F test was significant, F(3, 69) = 23.12, p < 0.001.

A Duncan’s multiple range test reveals again that the only sig-nificant difference is between the mouse travel technique and each of the three other travel techniques.

Here, the mean travelled distance was between 6% and 8% high-er when using the keyboard, joystick or gamepad than when using the mouse. The effect size is illustrated inTable 2.

Again, we see that there is a large effect size between the mouse travel technique and each of the three other travel techniques.

Fig. 10illustrates the results for the four travel techniques for the number of collisions per travelled meter (average of the last three trials). An analysis of variance (ANOVA) with the pseudo-F test was significant, F(3, 69) = 9.89, p < 0.001.

A Duncan’s multiple range test reveals a significant difference between the mouse and the joystick as well as between the game-pad and the three other travel techniques.

Using the gamepad increases from 33% to 122% the number of collisions per travelled meter as compared to the use of the key-board, mouse or joystick. The use of the joystick increases the number of collisions per travelled meter by 67% as compared to the use of the mouse. The effect size is illustrated inTable 3, com-puted from the left column.

Fig. 11illustrates the results for the four travel techniques for the number of collisions per second (average of the last three

Fig. 8. Mean completion time.

Table 1

Effect size for the completion time between the four interfaces.

Effect size Mouse Joystick Gamepad Keyboard 1.22 0.24 0.11 Mouse 1.84 0.92 Joystick 0.06 236.55 220.10 233.44 236.33 215 220 225 230 235 240 245

Keyboard Mouse Joystick Gamepad

Mean Travelled Distance (m) with Standard Error Bars

Travel Technique

Effect of Travel Technique on Travelled Distance

Fig. 9. Mean travelled distance.

Table 2

Effect size for the travelled distance between the four travel techniques. Effect size Mouse Joystick Gamepad Keyboard 1.92 0.31 0.02 Mouse 2.11 1.57 Joystick 0.25 1.23 0.94 1.57 2.09 0 0.5 1 1.5 2 2.5 3

Mean Number of Collision per

Travelled Meter

with Standard Error Bars

Effect of Travel Technique on the Number of Collision per Travelled Meter

Fig. 10. Mean number of collisions per travelled meter.

Table 3

Effect size for the number of collision per travelled meter for the four travel techniques.

Effect size Mouse Joystick Gamepad Keyboard 0.27 0.31 0.68

Mouse 0.52 1.17

(7)

Author's personal copy

trials). An analysis of variance (ANOVA) with the pseudo-F test was significant, F(3, 69) = 9.18, p < 0.001.

A Duncan’ multiple range test reveals a significant difference between the mouse and the joystick as well as between the game-pad and each of the three other travel techniques.

Using the gamepad interface increases from 38% to 95% the number of collisions per second as compared to the use of the key-board, mouse or joystick. The use of the joystick increases the number of collisions per second by 41% as compared to the use of the mouse. The effect size is illustrated inTable 4, computed from the left column.

Fig. 12reports the subjective ratings on a scale from 1 to 5, a higher score meaning a better score. Participants had to rate each interface according to five dimensions which are ease-of-use, fati-gue, accuracy, speed and preference.

Participants rated the mouse-based travel technique as their preferred method for virtual walkthroughs, overall, for ease of use, for accuracy, and for speed.

The results also indicate that the joystick was the least appreci-ated of the four travel techniques. Because of non-linearity, the subjective data were not subjected to statistical analysis.

4. Discussion

The results of this experiment indicate that, for this specific vir-tual environment, the mouse travel technique offers better perfor-mance than the three other travel techniques tested, namely the keyboard, the joystick and the gamepad travel techniques.

Accordingly, the subjective ratings indicate that the mouse tra-vel technique also provides a better user experience in regard to ease of use, speed, accuracy and overall preference.

It is important to note that the mouse travel technique uses its two buttons thus limiting the possibilities of further interaction with the environment (e.g. selecting an object with a click).

Given the popularity of the gamepad, these results can be sur-prising. There is however no contradiction as the gamepad does not require a desktop surface to operate, contrary to the three other input devices. It is therefore appropriate for users of video games who are generally located on a couch in front of a television. It could be interesting to see whether assigning the translations to another input device while conserving the mouse to control the orientation would offer better results. Given the different nature of translation (move) and rotation (look) actions, the idea of separat-ing those actions on different input devices could lead to better performance as indicated by previous studies (Jacob, Sibert, McFarlance, & Mullen, 1994). In fact, results of a previous experi-ment conducted with the same virtual environexperi-ment (Lapointe & Savard, 2009) indicates that using a bimanual travel technique made of a joystick and a mouse results in a much better perfor-mance than using a joystick only but slightly worse than using the single-handed mouse travel technique used here. This would hold true even if a third degree-of-freedom (strafe) is added as found in that previous experiment. In order to maximize perfor-mance, and given the results obtained here, it would be interesting to test a two-handed travel technique made of a keyboard (for move control) and a mouse (for look control). This travel technique is used extensively in first-person (egocentric) 3D computer games although no formal experiment demonstrated that this configura-tion offers the best usability for desktop virtual walkthroughs.

Overall it seems that the mouse has an advantage in terms of performance, but can quickly become overloaded if additional input channels are needed, for example to select objects in the vir-tual environment.

The results reported here are limited to one example of virtual environment and one particular set of users. Indeed, fully half of

8.03 7.23 10.22 14.12 0 2 4 6 8 10 12 14 16 18 20

Mean Number of Collisions per Second

with Standard Error Bars

Effect of Travel Technique on the Number of Collisions per Second

Fig. 11. Mean number of collisions per second.

Table 4

Effect size for the number of collisions per second for the four travel techniques. Effect size Mouse Joystick Gamepad Keyboard 0.12 0.34 0.81

Mouse 0.60 1.10

Joystick 0.66

Fig. 12. Subjective ratings.

(8)

our participants almost never played computer/video games. Only 2 of the 24 participants played almost daily. In contrast, all partic-ipants used a mouse on a daily basis. Consequently, we can state that our participants, as a whole, had much more experience with the mouse than with the joystick or gamepad. This is not surpris-ing, given that he mouse is a standard input device for desktop computers, and consequently, it is used for many other tasks be-sides game playing and virtual walkthroughs. This experience could have played a role in the mouse’s outperformance, but seem to reflect the experience of computers users in general.

For more comprehensive results, additional experiments should be conducted, by using tasks and virtual environments of different complexities with different sets of users. For example, tasks involv-ing walkthroughs in several different environments such as a straight line (1D), a plane (2D), or space (3D) could help show which DOF of the control interfaces are well suited to specific types of movement.

One important contribution of this paper is to report not only the results of this particular experiment, but also the evaluation method that was used and that could be reused for further exper-iments, including replications.

5. Conclusion

In conclusion, the mouse input device seems well suited as a single-handed device for simple walkthrough tasks in a desktop virtual environment.

Moreover, further experimentation is required to confirm the superiority of the mouse in other virtual environments and for other users.

6. Future work

Given the superiority of the mouse and previous findings show-ing high performance levels for bimanual techniques, travel could potentially be further improved by combining the mouse with a key-board as is often the case in first-person (egocentric) 3D computer games. Of course, this hypothesis awaits empirical confirmation.

Another future research direction could look at other input de-vices and interaction techniques such as those provided by sensor-based input techniques such as the Wii and Kinect.

Also, since standard testing methods do not exist in the field, more work in that direction would be needed in order to make sys-tematic progress in improving interaction techniques.

Acknowledgements

The authors thank all the volunteers who participated in this study. This experiment has been approved by the research ethics board from the National Research Council of Canada. This research has been funded by the Institute for Information Technology of the National Research Council of Canada.

References

Bowman, D. A., Kruijff, E., LaViola, J. J., Jr., & Poupyrev, I. (2005). 3D user interfaces: Theory and practice. Pearsons Education.

Brooks, F. P. (1986). Walkthrough – a dynamic graphics system for simulating virtual buildings. In Proceedings of workshop on interactive 3D graphics (pp. 9– 21).

Cohen, J. (1988). Statistical power analysis for the behavioral sciences (2nd ed.). Lawrence Erlbaum Associates, Inc.

Duncan, D. B. (1955). Multiple range and multiple F tests. Biometrics, 11, 1–42. Fröhlich, B., Hochstrate, J., Kulik, V., & Huckauf, A. (2006). On 3D input devices. IEEE

Computer Graphics and Applications, 26(2), 15–19.

Jacob, R. J. K., Sibert, L. E., McFarlance, D. C., & Mullen, M. P. Jr., (1994). Integrality and separability of input devices. ACM Transaction on Computer–Human Interaction, 1(1), 3–26.

Lampton, D. R., Knerr, B. W., Goldberg, S. L., Bliss, J. P., Moshell, J. M., & Blau, B. S. (1994). The virtual environment performance assessment battery (VEPAB): Development and evaluation. Presence: Teleoperators and Virtual Environments, 3(2), 145–157.

Lapointe, J. -F., Savard, P. (2007). Comparison of viewpoint orientation techniques for desktop virtual walkthroughs. In Proceedings of the IEEE international workshop on haptic audio visual environments and their applications (HAVE 2007) (pp. 33–37).

Lapointe, J. -F., Savard, P. (2009). A comparative study of three bimanual travel techniques for desktop virtual walkthroughs. In Proceedings of the IEEE international workshop on haptic audio visual environments and games (HAVE 2009) (pp. 182–185).

Lapointe, J. -F., Vinson, N. G. (2002). Effects of joystick mapping and field-of-view on human performance in virtual walkthroughs. In Proceedings of the IEEE international symposium on 3D data processing visualization and transmission (3D PVT 2002) (pp. 490–493).

Lindeman, R. W. (2006). Making VR more usable: The state of effectiveness in virtual reality. Journal of the Virtual Reality Society of Japan, 11(1), 24–27. Newell, A., & Card, S. K. (1985). The prospects for psychological science in human–

computer interaction. Human–Computer Interaction, 1(3), 209–242.

Ruddle, R. (1999). Navigating large-scale virtual environments: What differences occur between helmet-mounted and desk-top displays. Presence: Teleoperators and Virtual Environments, 8(2), 157–168.

Ruddle, R. A., Payne, S. J., & Jones, D. M. (1997). Navigating buildings in ‘‘desk-top’’ virtual environments: Experimental investigations using extended navigational experience. Journal of Experimental Psychology: Applied, 3(2), 143–159. Ruddle, R. A., Payne, S. J., & Jones, D. M. (1998). Navigating large-scale ‘‘desk-top’’

virtual buildings: Effects of orientation aids and familiarity. Presence: Teleoperators and Virtual Environments, 7(2), 179–192.

Santos, B. S., Dias, P., Pimentel, A., Baggerman, J. W., Ferreira, C., Silva, S., et al. (2009). Head-mounted display versus desktop for 3D navigation in virtual reality: a user study. Multimedia Tools Application, 41, 161–181.

Sebok, A., Nystad, E., & Helgar, S. (2004). Navigation in desktop virtual environments: an evaluation and recommendations for supporting usability. Virtual Reality, 8(1), 26–40.

Usoh, M., Arthur, K., Whitton, M., Steed, A., Slater, M., Brooks, F. (1999). Walking > virtual walking > flying, in virtual environments. In Proceedings of the ACM international conference on computer graphics and interactive techniques (SIGGRAPH 1999) (pp. 359–364).

Zhai, S. (1993). Investigation of feel for 6 DOF inputs: isometric and elastic rate control for manipulation in 3D Environments. In Proceedings of the human factors and ergonomics society annual meeting (HFES 1993) (pp. 323–327). Zhai, S. (1998). User performance in relation to 3D input device. In Proceedings of the

ACM international conference on computer graphics and interactive techniques (SIGGRAPH 1998) (pp. 50–54).

Zhai, S., Kandogan, E., Smith, B. A., & Selker, T. (1999). In search of the ‘magic carpet’: Design & experimentation of a bimanual 3D navigation interface. Journal of Visual Languages and Computing, 10(1), 3–17.